IRC channel logs

2024-10-04.log

back to list of logs

<mwette>What's the fix when slow builds of guile fail on guile-procedures.txt?
<rlb>mwette: depending on the failure, I just committed a fix for one problem 1c96e4ab6dde18c69f1493a8e1560e80a347cd21
<rlb>wingo: after a while it randomly occurred to me "could that one remaining r4rs.test ttyname failure (trying to print the backtrace -- no idea why) on powerpc" be due to the chroot and/or tmux I'm using on the porterbox. Sure enough, without tmux, all the tests pass now.
<rlb>wingo: so I think 3.0.11's good to go at least wrt the 32-bit issues. Though it'd be nice to add the x87 fix too (not critical). If you have something in mind there, great, otherwise I might attempt it, but don't let that hold up the release whenever you're ready.
<mwette>rlb: thanks, ill check tomorrow
<dialup>what is the best way for writing CSS inside Guile? Just writing CSS directly? I have currently written a function to generate "hyperscript"'s syntax to add interactivity directly in SXML, was curious if there was a macro/project for generating CSS
<tohoyn>sneek: botsnack
<sneek>:)
<mwette>ArneBab: wrt info patch, building on main branch now; had issues with the releases and guile-procedures.txt; not there but working it
<mwette>it hangs in 00-repl-server.test
<ArneBab>mwette: thank you! The repl-server.test should be fixed by -j2 (having two threads)
<mwette>ArneBab: really? I think the test spawns its own threads. I'm using `check-guile' also, it does not handle multiple threads.
<mwette>ArneBab: I think I found the issue. Not sure how to solve. The patch mostly replaces warning-port w/ info-port for the auto-compile calls etc. But it does also change the port to info-port from current-port for the initial help message. the repl-server test is looking for the end of the help message (from output port, aka /dev/1), but it's not going to be there since it's being pushed to the info port, aka /dev/2.
<mwette>There are a few options: 1) keep initial help message; 2) add command line argument to assign info to output versus error; 3) update repl-server-test (not staightforward)
<mwette>ArneBab: I think I will just keep the initial help message. For applications where guile is embedded it doesn't make sense (to me) to have my tool come up and say "GNU Guile" when it's not guile but for that I will write my own repl. I'll go for option (1). Let me know what you think.
<mwette>ArneBab: I have emailed an updated patch to the devel list.
<graywolf>Hello :) Yet another weird question: Is it possible do disable GC for a specific section of the code?